The overall goal of this proposal is to try to identify genes responsible for determining cell type specificity and eventually to try to understand the genetic complexity and logic of how this specificity is generated. We will be focusing on four types of developmental systems: early decision making in specific frog lineages; genes that commit cells to myogenesis; genes that determine whether hematopoietic precursor cells differentiate into macrophages or granulocytes; and genes that seem to link developmental decisions to decisions of growth control. For many of these systems, we begin by using cDNA clones that are cell type or cell lineage specific and we try to elicit a "phenotype" by introducing these sequences back into cells -- in a positive sense, to induce commitment with normal expression of the vector; in a negative sense, to inhibit differentiation with a vector that produces "abnormal" transcripts. We are hoping to create recombinant DNA mutations using 4 approaches. All of these methods must yield dominant mutations to be useful. The first is to use anti-sense RNA; the second is to use standard in vitro mutagenesis of cloned DNA to create a dominant mutation; the third is to overproduce the gene product; the fourth is to express a tissue-specific gene in the wrong cell type. For all of these schemes we will restrict our attention to tissue or lineage specific sequences obtained by "substractive" cloning of the unique RNAs (or cDNAs) peculiar to a given cell type. In most cases, this material is readily available. Estimates from solution hybridization suggest that if the two cell types that are to be compared are closely related, one might expect between 200-1000 sequences in such a difference library. These could be screened either in pools or individually. In the case of red blood cells, the screen is to look for embryos that lack red cells.